INTRODUCTION
Despite family firms being the most common economic organization, nowadays research dealing with its prevalence and its distinctive characteristics remains scarce (Schulze & Gedajlovic, Reference Schulze and Gedajlovic2010). More specifically, the study of top management teams (TMTs) attracts very limited attention, with few exceptions being the work carried out by Ensley and Pearson (Reference Ensley and Pearson2005) or, more recently, works by Minichilli, Corbetta, and MacMillan (Reference Minichilli, Corbetta and MacMillan2010) and Ling and Kellermanns (Reference Ling and Kellermanns2010). This absence in the literature is significant, as management teams are a fundamental element in the governance of family firms (Neubauer & Lank, Reference Neubauer and Lank1998). In fact, in these firms, particularly non-listed ones, management teams are the ones that really influence the strategic orientation and thus the firm's performance. This usually occurs because the board of directors, when exists, is an underutilized instrument, which fulfills a purely legal requisite and that, in many cases, limits itself to endorse the decisions already made by the management team (Dyer, Reference Dyer1986; Ward, Reference Ward1991).
General literature on TMTs highlights their composition and structure, as well as their behavioral dynamics, as relevant variables to explain their efficiency. In this way, the literature on TMTs is fundamentally based on the upper echelon theory approach and the theory regarding process and group dynamics (for a review, see, e.g., Carpenter, Geletkanycz, & Sanders, Reference Carpenter, Geletkanycz and Sanders2004; Certo, Lester, Dalton, & Dalton, Reference Certo, Lester, Dalton and Dalton2006).
In general literature, the composition of TMTs is explained in terms of age difference, gender, seniority, or functional and educational backgrounds of the team members (Certo et al., Reference Certo, Lester, Dalton and Dalton2006). However, in family firms, one of the most relevant issues is whether the members of the TMT belong or not to the owning family (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010). Specifically, one of the main characteristics of non-listed family firms is the high need for control, which results in the owning family having an almost absolute power over the ownership of the firm, as well as over the board of directors (Dyer, Reference Dyer1986; Neubauer & Lank, Reference Neubauer and Lank1998; Cabrera-Suárez & Santana-Martin, Reference Cabrera-Suárez and Santana-Martin2004). Thus, a way of adding a certain degree of diversity in strategic decision making is the inclusion of non-family members to the management teams (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010). On the other hand, the family nature of the top executive of the firm (chief executive officer or general manager) is an important variable in the analysis of the management team composition, due to the strong influence such a figure has on decision making, particularly in family firms (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010). Likewise, in the specific case of top family executives, this paper considers that the presence of the founder in management teams will give rise to differences in the decision-making process. In fact, the figure of the founder is specially relevant in family firms due to the special characteristics that distinguish them in terms of commitment and knowledge of the firm, personality and leadership style, etc. (Dyer, Reference Dyer1986, Reference Dyer2006; Kets de Vries, Reference Kets de Vries1993; Sorenson, Reference Sorenson2000; Ling & Kellermanns, Reference Ling and Kellermanns2010).
Variation within family managers can also be an important factor related to the composition of management teams of family firms. This last variable has barely been discussed in research on TMTs in family firms, one of the few exceptions being the work by Ensley and Pearson (Reference Ensley and Pearson2005). These authors distinguish between TMTs, where parental links exist and those with no such parental influence. This paper also analyzes this specific variable using as an indicator whether the family managers are siblings or not.
In relation to the behavioral dynamics of TMT, the literature indicates that the TMTs must achieve behavioral integration between its members, referring to the capability of a TMT to exchange information, behave in a collaborative manner, and carry out joint decision making (Simsek, Veiga, Lubatkin, & Dino, Reference Simsek, Veiga, Lubatkin and Dino2005; Lubatkin, Simsek, Ling, & Veiga, Reference Lubatkin, Simsek, Ling and Veiga2006; Buyl, Boone, Hendriks, & Matthyssens, Reference Buyl, Boone, Hendriks and Matthyssens2011). In fact, the ability of the members of a team to interact, worry about one another, communicate and generate feelings of belonging, trust and networks of shared meaning is the key to develop fundamental processes for the organizations such as the creation of knowledge (Styhre, Roth, & Ingelgard, Reference Styhre, Roth and Ingelgard2002).
This definition of behavioral integration, in the specific case of family firms, leads to highlight the need to include ideas related with the concept of familiness when analyzing their TMTs. Familiness is defined as a bundle of idiosyncratic resources and capabilities resulting from interactions between the family, the firm and individual family members (Habbershon, Williams, & MacMillan, Reference Habbershon, Williams and MacMillan2003). Recent developments in the concept of familiness tend to emphasize the relation of this concept to that of social capital (Arregle, Hitt, Sirmon, & Very, Reference Arregle, Hitt, Sirmon and Very2007; Pearson, Carr, & Shaw, Reference Pearson, Carr and Shaw2008; Sharma, Reference Sharma2008). Thus, we state that the relation between the elements of social capital (specifically its cognitive and relational dimension), associated with familiness and behavioral integration, is fundamental for understanding the effective functioning of TMTs in family firms.
Therefore, this paper, based on this approach on familiness, will develop the theoretical arguments on how the composition and the structure of management teams in family firms affect their behavior and therefore the performance of the firms. In this sense, the specific context of non-listed family firms is especially suitable given that it is assumed that private family firms are more likely to show the features that create familiness. In addition, this paper contributes to the study of family firms following the lines of investigation suggested by researchers such as Ling and Kellermanns (Reference Ling and Kellermanns2010). They establish the need to advance the study of TMTs of family firms, determining the real composition of these teams and distinguishing between the effects of diversity induced by the family and those induced by non-family members. With this aim, a database including several variables related to the composition of management teams of Spanish non-listed family firms has been created.
The structure of the study is as follows. The following section includes a theoretical argumentation that discusses the specific nature of private family firms in terms of familiness and social capital, followed by the arguments that support the hypotheses in relation to the effect of the composition of TMTs on performance. The third section outlines the methodology used to test hypotheses and present the results. The final section offers the discussion of the results and the main conclusions, as well as their implications and suggestions for future research.
THEORETICAL FRAMEWORK
The private family firm and familiness
The family nature of a business is determined by the cultural and behavioral aspects introduced by the controlling family whose members strive to achieve and/or maintain intra-organizational family-based relationships. Thus, an essence approach to the concept of family firm can be adopted (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma1999; Chrisman, Chua, & Sharma, Reference Chrisman, Chua and Sharma2005; Zellweger, Eddleston, & Kellermanns, Reference Zellweger, Eddleston and Kellermanns2010; Gomez-Mejia, Cruz, Berrone, & De Castro, Reference Gomez-Mejia, Cruz, Berrone and de Castro2011). Our line of argument, and consequently our hypotheses, is based on the premise that private family firms possess more defining characteristics of the essence of family firms than public family firms. Thus, private family firms correspond to what the literature considers typical family firms, with a concentrated shareholder base and family member insiders active in management and the board (Lane, Astrachan, Keyt, & McMillan, Reference Lane, Astrachan, Keyt and McMillan2006).
Privately held family businesses are often used as vehicles for sustaining the family's transgenerational economic and socio-emotional needs (Bammens, Voordeckers, & Van Gils, Reference Bammens, Voordeckers and Van Gils2011). Families are actively involved not only in ownership, but also in the management of the firm (Westhead & Howorth, Reference Westhead and Howorth2006; Sciascia & Mazzola, Reference Sciascia and Mazzola2008). The existence of non-family shareholders is very unlikely and there is a strong presence of family members in management teams and board of directors (Dyer, Reference Dyer1986; Cabrera-Suárez & Santana-Martin, Reference Cabrera-Suárez and Santana-Martin2004; Lubatkin, Schulze, Ling, & Dino, Reference Lubatkin, Schulze, Ling and Dino2005; Lane et al., Reference Lane, Astrachan, Keyt and McMillan2006). Thus, family ties and relationships have a great influence in the functioning of the firm (Miller, Le Breton-Miller, Lester, & Cannella, Reference Miller, Le Breton-Miller, Lester and Cannella2007). On the contrary, listed firms will include non-family owners that will want to play an active role in the governance of the firms and in decision-making process, probably wanting to impose economic and financial criteria over psychosocial criteria. Listed family firms are subjected to other pressures and conditioning factors over and above fulfilling the interests, aspirations and the approach of the family (Dyer, Reference Dyer1986; Combs, Reference Combs2008; Bammens, Voordeckers, & Van Gils, Reference Bammens, Voordeckers and Van Gils2011).
Non-listed family firms are probably smaller than their listed counterparts, and this size difference is reflected both in the firm and the family. Smaller firms require less formality and sophistication in their management. On the other hand, large listed firms that are exposed to public scrutiny and the regulations of the stock markets will require a more professional management and will adopt behavior patterns different to those typical of family firms (Dyer, Reference Dyer1986; Neubauer & Lank, Reference Neubauer and Lank1998; Lane et al., Reference Lane, Astrachan, Keyt and McMillan2006). In addition, families in listed firms tend to be large and disperse, where family relationships are weak or even prone to conflict (Dyer, Reference Dyer1986; Gersick, Davis, McCollom-Hampton, & Lansberg, Reference Gersick, Davis, McCollom-Hampton and Lansberg1997).
Thus, it can be argued that private family firms are more likely to show the specific and distinctive nature of a family firm, that is ‘familiness’ (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma1999; Sirmon & Hitt, Reference Sirmon and Hitt2003). The concept of familiness was first defined by Habbershon and Williams (Reference Habbershon and Williams1999) as the bundle of idiosyncratic internal resources and capabilities resulting from the involvement of the family in the firm. The involvement of family members with the firm and their interactions within it has been put forward by several authors as the source of certain family-based attributes of family firms that create familiness (e.g., Chrisman, Chua, & Litz, Reference Chrisman, Chua and Litz2003; Sirmon & Hitt, Reference Sirmon and Hitt2003; Pearson, Carr, & Shaw, Reference Pearson, Carr and Shaw2008; Sharma, Reference Sharma2008; Zellweger, Eddleston, & Kellermanns, Reference Zellweger, Eddleston and Kellermanns2010). The theory of social capital is one of the most recent theoretical developments regarding the construct of familiness (Hoffman, Hoelscher, & Sorenson, Reference Hoffman, Hoelscher and Sorenson2006; Arregle et al., Reference Arregle, Hitt, Sirmon and Very2007; Pearson, Carr, & Shaw, Reference Pearson, Carr and Shaw2008; Sharma, Reference Sharma2008).
Three dimensions of social capital can be regarded as the specific elements of familiness (Sirmon & Hitt, Reference Sirmon and Hitt2003; Hoffman, Hoelscher, & Sorenson, Reference Hoffman, Hoelscher and Sorenson2006; Pearson, Carr, & Shaw, Reference Pearson, Carr and Shaw2008). First, the structural dimension refers to the internal network of ties inside a family resulting from the established patterns of interaction and involvement among relatives that can be appropriated by the firm. Second, the cognitive dimension refers to the family's shared vision and purpose, as well as its unique language, stories and culture that are known and understood, allowing shared communication and integration of ideas. Third, the relational dimension comprises the resources created through personal relationships such as trust, norms, obligations and identity, essential in the family firm to identify and pursue not only economic goals but also value creation goals specific to the family. The three dimensions of social capital that constitute familiness will depend on the strength of the family ties. These will depend on the possibility that family members have to interact and communicate. Thus, we expect families in large listed firms to present these characteristics in a smaller degree than owning families of closely held family firms.
This study considers that certain characteristics of the composition of TMTs in private family firms will affect the dynamics associated to familiness and therefore the behavior of TMT. These characteristics are: the proportion of non-family managers; the diversity in family managers considering the link between them (siblings vs. non-siblings); the family nature of the top executive and whether the founder is a member of the management team. The following paragraphs analyze each of these aspects and develop the argumentation for the hypothesized effect on the performance of non-listed family firms.
Family managers versus non-family managers
Several studies in the literature mention the benefits of including non-family members into the TMT. The main advantages are related to an increase in formalization, professionalization, and the knowledge and abilities of the management teams (Dyer, Reference Dyer2006; Blumentritt, Keyt, & Astrachan, Reference Blumentritt, Keyt and Astrachan2007; Chittoor & Das, Reference Chittoor and Das2007; Sciascia & Mazzola, Reference Sciascia and Mazzola2008). Professional managers selected in a competitive labor market, external to the family, are expected to show greater directive abilities (Sirmon & Hitt, Reference Sirmon and Hitt2003; Lin & Hu, Reference Lin and Hu2007). The inclusion of these non-family managers can increase the level of heterogeneity and professionalism in the TMT, encourage an analytical orientation in problem solving, as well as increase innovation, entrepreneurial capacity and the emerging of opposing ideas (Sirmon & Hitt, Reference Sirmon and Hitt2003; Zahra, Hayton, & Salvato, Reference Zahra, Hayton and Salvato2004; Escribá-Esteve, Sánchez-Peinado, & Sánchez-Peinado, Reference Escribá-Esteve, Sánchez-Peinado and Sánchez-Peinado2009). The inclusion of non-family managers could reduce the likelihood of incurring agency costs derived from problems of parental altruism, nepotism and adverse selection in the labor market. These problems entail the risk of hiring non-qualified personnel, and the perpetuation in directive roles of people no longer competent to run the firm (Lubatkin et al., Reference Lubatkin, Schulze, Ling and Dino2005; Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010; Voordeckers, Lambrechts, & Steijvers, Reference Voordeckers, Lambrechts and Steijvers2010).
However, for the specific case of non-listed family firms, including managers who do not form a part of the owning family also entails some potentially negative aspects. Thus, the increase in the proportion of non-family managers can lead to the gains in cognitive diversity and increased intellectual capital of the team (Sciascia & Mazzola, Reference Sciascia and Mazzola2008) being offset by the losses in terms of familiness, as the positive dynamics associated with teams composed of family members will have the tendency to disappear. The emotional attachment and identification of family managers to their firms (Miller & Le Breton-Miller, Reference Miller and Le Breton-Miller2006; Chrisman, Chua, Kellermans, & Chang, Reference Chrisman, Chua, Kellermanns and Chang2007; Jones, Makri, & Gómez-Mejía, Reference Jones, Makri and Gomez-Mejia2008; Cruz, Gómez-Mejía, & Becerra, Reference Cruz, Gómez-Mejia and Becerra2010) are probably less intense in non-family managers. Therefore, the expectation is that non-family managers will identify to a lesser degree with the culture, values and norms shared by the family members. Thus, non-family managers have less emotional and psychological elements that constitute the basis of reciprocal altruism and stewardship-type motivations (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma2003). Interdependence, values and shared norms are the main characteristics of reciprocal altruism (Lubatkin, Duran, & Ling, Reference Lubatkin, Durand and Ling2007). This type of altruism encourages feelings of loyalty and trust between family members, facilitates communication and lengthens the decision-making time frame, all of which ultimately reduce agency costs (Bartholomeusz & Tanewski, Reference Bartholomeusz and Tanewski2006; Karra, Tracey, & Phillips, Reference Karra, Tracey and Phillips2006; Lubatkin, Duran, & Ling, Reference Lubatkin, Durand and Ling2007; Sciascia & Mazzola, Reference Sciascia and Mazzola2008). Thus, if the proportion of non-family managers increases, decision making is less likely to be determined by objectives related to generational transmission of the firm, maintenance of reputation and relations with stakeholders, and other objectives associated with producing long-term value and reducing opportunistic behavior (Bartholomeusz & Tanewski, Reference Bartholomeusz and Tanewski2006; Eddleston & Kellermanns, Reference Eddleston and Kellermanns2007; Sciascia & Mazzola, Reference Sciascia and Mazzola2008; Thomas, Reference Thomas2009; Kowalewski, Talavera, & Stetsyuk, Reference Kowalewski, Talavera and Stetsyuk2010; Schulze & Gedajlovic, Reference Schulze and Gedajlovic2010). As a consequence, the increase of non-family managers could result in an increase of problems and traditional agency costs; as the family firm becomes more professional, including non-family managers and delegating more responsibility to them, the firm will resemble non-family firms in terms of divergence of interests between agents and principals (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma2003; Dyer, Reference Dyer2006; Voordeckers, Lambrechts, & Steijvers, Reference Voordeckers, Lambrechts and Steijvers2010). This, in turn, leads to problems of information asymmetries and limited rationality between the owners of the family firm and non-family managers.
On the other hand, non-family managers may share between them the feeling that they are excluded from the family that controls the firm. Consequently, the growing proportion of non-family managers can lead to what Minichilli, Corbetta, and MacMillan (Reference Minichilli, Corbetta and MacMillan2010) refer to as a faultline between the managers, which do and do not belong to the family. This faultline can produce disagreements and tensions between the two groups, negatively affecting the dynamics of behavior in the management teams and, as a consequence, the results.
In sum, the inclusion of non-family managers can bring about a valuable contribution to the management teams in family firms in terms of cognitive diversity, professionalism and objectivity. However, an increased proportion of non-family managers beyond a certain level can destroy the specific advantages of family firms regarding social capital associated with familiness. Therefore:
Hypothesis 1: An inverted U relationship exists between the number of non-family managers and performance of non-listed family firms.
Diversity of family managers
The discussion in the previous paragraphs implicitly states that family managers constitute a homogeneous group, with shared values, experiences and expectations regarding the family firm. However, the possibility of family managers having different aims and values could also be stated. Even when they belong to the same family, the more distant the parental tie and the dispersion of family members, the more likely beliefs and group ties will weaken (Ensley & Pearson, Reference Ensley and Pearson2005; Dyer, Reference Dyer2006). Ling and Kellermanns (Reference Ling and Kellermanns2010) highlight that homogeneity between family members who work in the family firm should not be assumed. Members will differ according to the different education and socialization processes they have undergone due to different experiences and social networks. However, differences in the development processes of family members are probably less between siblings from the same mother and father and who have lived in the same household, than between cousins or even parents and children. In this last case, differences in attitudes and values linked to the age gap and their different motivations and experiences will be higher.
These differences can negatively affect familiness. Certain aspects such as trust, a sense of mutual obligation, norms of cooperation, identification with the family group and the firm, which constitute the relational dimension of social capital, could weaken as family ties lessen (Salvato & Melin, Reference Salvato and Melin2008). Reduced family cohesion can make the firm more susceptible to experiencing destructive conflicts capable of dividing the family into factions based on generations or family branches, which detracts from performance (Björnberg & Nicholson, Reference Björnberg and Nicholson2007; Kellermanns & Eddleston, Reference Kellermanns and Eddleston2007; Miller, Le Breton-Miller, & Scholnick, Reference Miller, Le Breton-Miller and Scholnick2008).
Therefore, the higher the proportion of siblings relative to the total number of relatives in management teams (greater family homogeneity), the greater the probability that these management teams show the positive aspects associated with familiness in terms of social capital. Similarly, the lower the proportion of siblings (greater heterogeneity), the greater the impact on factors that improve the decision-making process (e.g., frequency of information exchange, trust, cognitive conflict), and the factors that hinder this process (e.g., affective conflict) increases. Thus:
Hypothesis 2: A positive relation exists between homogeneity of family managers and non-listed family firm performance.
The family nature of the top executive
The behavior and efficiency of TMTs depends on various factors related to their composition and the characteristics of their members. However, the literature specifically admits the influence of the team leader as a determining factor on their performance (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010; Buyl et al., Reference Buyl, Boone, Hendriks and Matthyssens2011). The behavior of executive leaders is crucial to the improvement of certain attitudes of their subordinate managers (Ensley & Pearson, Reference Ensley and Pearson2005; Schulze & Gedajlovic, Reference Schulze and Gedajlovic2010) such as organizational commitment, which ultimately has an influence on the firm's results. In family firms the leader's influence on the management team increases if the person belongs to the owning family (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010).
Family leaders of private family firms possess certain characteristics and display behaviors that can be particularly beneficial. The strong ties characteristic of family social capital lead to extensive closure, which provides for collective sanctions if members of the family business do not fulfill expectations in relation to their conduct both within the business and outside the network (Hoffman, Hoelscher, & Sorenson, Reference Hoffman, Hoelscher and Sorenson2006). This encourages the existence of ethical norms shared by family members that can improve both relationships with stakeholders and the reputation of the firm (Sorenson, Goodpaster, Hedberg, & Yu, Reference Sorenson, Goodpaster, Hedberg and Yu2009), adopting a steward-like perspective (Le Breton-Miller & Miller, Reference Le Breton-Miller and Miller2006; Chen & Hsu, Reference Chen and Hsu2009). The main assumptions of stewardship theory (Davis, Schoorman, & Donaldson, Reference Davis, Schoorman and Donaldson1997) support the idea that executive behavior may be motivated by the general interests of a firm and not only by the private interests of executives. These assumptions are based on a conceptualization of the ‘organization man’ as motivated by achievement, responsibility and recognition, and who benefits more from working for the organization's interest than against it (Braun & Sharma, Reference Braun and Sharma2007).
Similarly, family managers experience a feeling of ongoing commitment to their firms, which is strengthened by the fact that they hold long tenures in their positions (Miller & Le Breton-Miller, Reference Miller and Le Breton-Miller2006). This behavior can result in the development of specific advantages related to greater opportunities for learning, and acquiring specific knowledge about their firms (Cabrera, De Saá, & García 2001; Miller & Le Breton-Miller, Reference Miller and Le Breton-Miller2006; Sacristan-Navarro & Gómez-Ansón, Reference Sacristán-Navarro and Gómez-Ansón2009). Thus:
Hypothesis 3: A positive relation exists between the family nature of the leader of TMTs and non-listed family firm performance.
However, the presence of a top family executive could entail certain negative aspects regarding their influence on non-family managers and their contribution to the management team. Minichilli, Corbetta, and MacMillan (Reference Minichilli, Corbetta and MacMillan2010) state that a family leader in the TMT could increase the division between relatives and non-family members in the management team and have a negative effect on group dynamics. According to these authors, non-family managers could feel more excluded and less identified with the firm in those situations and may have less opportunities of promotion. The strength of the ties with the CEO is one of the most significant factors when determining the opportunities of promotion of managers (Kim & Cannella, Reference Kim and Cannella2008). In family firms the logical assumption is that family executives will have the strongest ties with a family CEO and thus non-family executives may feel they have fewer opportunities for promotion. Thus, if non-family managers perceive an unfair treatment, they can reduce their level of commitment, satisfaction and cooperation (Barnett & Kellermanns, Reference Barnett and Kellermanns2006). The perception of exclusion may lead to feelings of alienation and the adoption of passive behavior (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma2003; Barnett & Kellermanns, Reference Barnett and Kellermanns2006). Therefore:
Hypothesis 4: When the leader of the TMT is a member of the family, the increase in the proportion of non-family managers will have a negative effect on the performance of non-listed family firms.
The presence of the founder in the management team
The figure of the founder is especially relevant in TMTs in general (Buyl et al., Reference Buyl, Boone, Hendriks and Matthyssens2011) and in family firms in particular (Kets de Vries, Reference Kets de Vries1985, Reference Kets de Vries1996; Kelly, Athanassiou, & Crittenden, Reference Kelly, Athanassiou and Crittenden2000; Dyer, Reference Dyer2006). In fact, the strength of values, objectives and intentions imposed by the founder (Dyer, Reference Dyer1986; Gedajlovic, Lubatkin, & Schulze, Reference Gedajlovic, Lubatkin and Schulze2004), his/her knowledge and specific abilities (Cabrera-Suárez, Saá-Pérez, & García-Almeida, Reference Cabrera-Suárez, Saá-Pérez and García-Almeida2001; Gedajlovic, Lubatkin, & Schulze, Reference Gedajlovic, Lubatkin and Schulze2004; Kowalewski, Talavera, & Stetsyuk, Reference Kowalewski, Talavera and Stetsyuk2010) as well as his/her high level of commitment (Ling & Kellermanns, Reference Ling and Kellermanns2010) constitutes a strategic resource in the decision-making teams of family firms.
The positive aspects of familiness will be stronger during the phase of the founders. Ensley and Pearson (Reference Ensley and Pearson2005), in their study on management teams in recently established firms, reach the conclusion that the TMTs they call ‘parental’ show more efficient behavior dynamics than those teams referred to as ‘familial’, constituted by siblings and/or cousins and where the ‘parental’ figure is no longer present. According to these authors, parental teams are similar to the category of controlling owner by Gersick et al. (Reference Gersick, Davis, McCollom-Hampton and Lansberg1997), in other words, they are similar to teams where the figure of the founder is present. Compared to familial teams, parental teams show a stronger belief in the team, stronger cohesion ties, greater agreement in their strategic vision and less relationship conflicts. Moreover, the special leadership characteristics of the founders can generate feelings of psychological ownership, stewardship motivations and reciprocal altruism not only between family members but also between non-family members who become quasi-relatives (Dyer, Reference Dyer2006; Karra, Tracey, & Phillips, Reference Karra, Tracey and Phillips2006).Thus:
Hypothesis 5a: A positive relation exists between the presence of the founder in TMTs and non-listed family firms performance.
However, other arguments in the literature highlight the potentially negative effects of the presence of the founder as a member of the TMT. The founders hold virtually all authority in their firms, as they tend to be the sole owners of the firm and thus their decisions are legitimized by their ownership rights (Gedajlovic, Lubatkin, & Schulze, Reference Gedajlovic, Lubatkin and Schulze2004; Harris & Ogbonna, Reference Harris and Ogbonna2007). The accumulation of power in this person makes them more dominant and less susceptible to consider contributions of other members of management teams, which in turn negatively affects expressing ideas and exchanging information (Buyl et al., Reference Buyl, Boone, Hendriks and Matthyssens2011). In parental TMTs, where the founder is present, the degree of conflict of ideas is lower (Ensley & Pearson, Reference Ensley and Pearson2005). According to these authors, the parental figure has great strength as a dominant leader and this can lead them to be reluctant to accept the ideas and suggestions of other relatives or lead to hesitation by members of the family to express ideas that question the patriarch's point of view. The governance structure of firms run by their founders makes firms vulnerable to problems of self-control that expose the firm's stakeholders to expropriation problems. This could lead stakeholders, which include other managers, to avoid making specific investments in these firms and to limit their contributions in terms of efforts or valuable information (Gedajlovic, Lubatkin, & Schulze, Reference Gedajlovic, Lubatkin and Schulze2004). Thus:
Hypothesis 5b: A negative relation exists between the presence of the founder in TMTs and non-listed family firms performance.
METHOD
Data
Non-listed Spanish firms are the population of this study. In this sense, one of the main obstacles in the study of non-listed family firms is the lack of reliable and systematic data on this type of firm (Uhlaner, Wright, & Huse, Reference Uhlaner, Wright and Huse2007; Arosa, Iturralde, & Maseda, Reference Arosa, Iturralde and Maseda2010). In Spain there is no official database of non-listed family firms, thus we had to create this database by indirectly identifying these firms from a database provided by Informa Dun and Bradstreet. This company was asked to list all the firms on its database, whose board of directors and/or management team included a minimum of two individuals with different first names who had the same two surnames in common. As all individuals in Spain receive two surnames, one from each parent, so people whose two surnames are identical are very likely to be siblings. The database was made up of 4,217 firms, which were potential family firms due to the presence of siblings in their TMT. After refining and analyzing the available information for each of the firms and completing with secondary sources, the following relative variables were added to the composition of the management team.
• The size of the management team and the relation between the positions with responsibility, including that of the chief executive.
• Number of siblings in the management team, as assumed when the two surnames coincide but the individuals have different first names. These two surnames are later used as a reference to identify other family members involved in the firm.
• Number of family members in the management team, as assumed when one of the previously identified surnames coincides.
• The family nature or not of the chairman of the board of directors.
• The family nature or not of the chief executive and/or general manager.
Information of these variables could be obtained for 2,179 firms. Finally, to ensure that the characteristics of the companies would enable the objectives of the study to be achieved, firms in which the following conditions existed were selected:
• The chairman of the board of directors belongs to the family.
• The management team has at least three positions of responsibility.
• The number of employees is equal to or more than 10. The main reason for only selecting firms with more than 10 employees was to exclude micro-firms. Thus, firms with more than 10 employees are more complex and the members of the management team are more likely to have different responsibilities and specializations. Therefore, the importance of behavioral integration between them is higher.
• The firm does not belong to the financial sector nor is listed on the stock market.
Consequently, we consider a firm to be a family firm if it has at least two people with different first names and two identical surnames (i.e., they are siblings) on the board of directors and/or management teams, and the chairman of the board has at least one of these two surnames (i.e., he/she is a family member). That is how we tried to ensure that the identified firms are in essence family firms. On the one hand, these two criteria ensure there is a real family influence on decision making, which is an essential form of family involvement that shapes the distinctiveness of a family firm (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma1999; Fiegener, Reference Fiegener2010). On the other hand, the presence of siblings in the governing bodies implies an intention, or in fact, the transmission of leadership between generations in the family; this, in turn, is another key factor in the definition of family firms (Chua, Chrisman, & Sharma, Reference Chua, Chrisman and Sharma1999; Astrachan, Klein, & Smyrnios, Reference Astrachan, Klein and Smyrnios2002; Sirmon & Hitt, Reference Sirmon and Hitt2003; Zellweger, Eddleston, & Kellermanns, Reference Zellweger, Eddleston and Kellermanns2010). Finally, the presence of family members on the board and the fact that the chair of the board is a family member also allows us to infer that the ownership of the firm is in fact controlled by the family (García-Castro & Casasola, Reference García-Castro and Casasola2011). Previous studies on property and control structures of public family firms in Spain show that in more than two-thirds of the firms that have an individual or a family as ultimate owner, a member of the family holds the post of CEO or chairman of the board (Sacristán-Navarro & Gómez-Ansón, Reference Sacristán-Navarro and Gómez-Ansón2007). If this is the case in large listed firms, we can expect this to occur in non-listed family firms.
Thus, 929 firms were finally included in the study. Their general characteristics can be found in Table 1. Over half of the firms (54.9%) are between 11 and 30 years old, while over a third (36.7%) was over 30. Most of the firms are public limited companies. Almost 60% of the firms were involved in the industrial sector, while 40% were in the service sector. In 52.3% of the firms, the employees numbered between 10 and 100 people, whilst ∼20% of the firms had more than 200 people working for them. Turnover in 2007 varied between firms, with 43.9% of them having a turnover of between 10 and 30 million euros and 29.8% of the firms more than 30 million. Finally, the majority (75.1%) had a management team with three or four members.
Table 1 Profile of analyzed firms
Variables
Dependent Variable
The natural logarithm of the ratio of sales to employees in 2007 was used as the outcome variable and, therefore, as the dependent variable. This is a productivity indicator that allows the effect of firm size on sales figures to be relativized. The ratio of productivity to sales was used because, in non-listed family firms, performance measures related to sales are considered to be more reliable than other account items subject to taxation (Schulze, Lubatkin, Dino, & Buchholtz, Reference Schulze, Lubatkin, Dino and Buchholtz2001; Oswald, Muse, & Rutherford, Reference Oswald, Muse and Rutherford2009). The logarithmic transformation was applied to make the variable behave according to a normal distribution (Kolmogorov–Smirnov Z = 3.813, p = .000 for the initial variable; Kolmogorov–Smirnov Z = 1.031, p = .238 for the transformed variable).
Independent variables
In order to satisfy the requirements of the hypotheses the following variables were created:
• The proportion of non-family members in the management team (PNFM), calculated as the quotient between the number of non-family members in the management team and the total size of the management team.
• The homogeneity of the family managers (HOMOGFM), which has been obtained calculating the quotient between the number of siblings (calculated by a match in both surnames) and the total number of family managers (at least one surname coincides) in the management team.
• The family nature of the head of the TMT (HEADFAM) is a dichotomous variable, with a value of 1 when the general manager or the CEO are family members, and 0 when this is not the case. In the sample studied, 85.8% of the firms had a family member as the head of the TMT.
• The presence of the founder in the management team (FOUNTMT). This dichotomous variable has a value of 1 when the general manager or the chief executive is a member of the owning family and is not one of the identified siblings, and the firm is less than 30 years of age. With this set of criteria the assumption is that the firm has not gone past its first succession, and that the top executive is the founder. In the samples analyzed, the founder formed part of the management team in 14.6% of the firms.
Moderating Effect
The moderating effect in Hypothesis 4 is linked to an interaction variable that tests whether the head of the management team being a family member affects the effect that the proportion of non-family managers has on the performance of the firm. Therefore, the moderating variable corresponds with the interaction of the independent variables PNFM and HEADFAM and is referred to it as PNFM × HEADFAM.
Control Variables
The following control variables have been included in the multiple linear regression model:
• Sector of activity: A dummy variable was created, using the tertiary sector as a reference. The variable DUMSEC has a value of ‘1’ if the firm belongs to the secondary sector and ‘0’ if this was otherwise. Consequently, the regression coefficients of the dummy variable represent the differential effect of the secondary sector on the dependent variable in relation with the reference category, which corresponds with the tertiary sector.
• The age of the firm: Measured as the number of years since the establishment of the firm. In order to minimize the asymmetry presented in the data, as was the case of the dependent variable, a natural logarithm was applied to the data.
RESULTS
Multicollinearity is one of the problems encountered in all multiple regression models, but particularly in moderated regressions. In order to determine the presence or absence of multicollinearity between the variables in the models the following tests were carried out: (1) the levels of correlation that exist between the variables of continuous nature, (2) the values of the variance inflation factor, and (3) the levels of tolerance.
Table 2 shows the correlation levels that exist between the continuous variables used in the regression model. Discriminant validity exists insofar as none of the correlation coefficients is 1, consequently, it is possible to assume that multicollinearity is not a problem. This same conclusion can be reached using the variance inflation factor values and the tolerance levels (Table 3) as the variance inflation factor values do not reach the threshold of 10 and the tolerance levels are higher than 0.10.
Table 2 Correlation coefficients between continuous variables in the model
Note. NLSALESEMPL07 = the natural logarithm of the ratio of sales to employees in 2007; HOMOGFM = homogeneity of the family managers; NLAGE = natural logarithm; PNFM = proportion of non-family members in the management team.
*, ** Significant correlation at p < .05 and p < .01, respectively.
Table 3 Results of regression models
Note. DUMSEC = sector of activity dummy variable; FOUNTMT = founder in the management team; HEADFAM = family nature of the head of the top management teams; HOMOGFM = homogeneity of the family managers; NLAGE = natural logarithm; PNFM = proportion of non-family members in the management team; VIF = variance inflation factor.
The hypotheses were tested using a hierarchical regression analysis, where the variables were introduced in successive blocks (Table 3). Model I analyses the effects of the control variables on the variable of the results to be analyzed. Model II incorporates the control variables and the independent variables. Model III only differs from Model II in the inclusion of the independent variable PNFM squared (PNFM2), in order to determine whether a quadratic relationship exists between the proportion of non-family managers and the performance of the firm. Finally, Model IV includes all the variables in Model II and the moderating effect. Moderated regression analysis was used to determine the effect of the moderating variable. This method is an appropriate way of studying whether or not moderator effects are present. The moderating effect is significant if the change observed in the coefficient of determination (R 2) is significant. The empirical evidence indicates that an increase of more than 1% can be considered significant, and consequently indicates the existence of a relevant moderating effect.
The results of Model I, found in the second column of Table 3, indicate that the sector of activity in which the firm operates is a determinant variable of the firm's results, with the firms in the secondary sector obtaining higher productivity. This result was expected, as firms in the service sector are, as a general rule more labor intensive, hence the ratio of productivity that has been employed as a dependent variable in this study will be negatively affected when compared to the tertiary sector. The firm's age has a positive effect on its performance, suggesting that older firms have a greater possibility of obtaining higher productivity.
The four explanatory variables considered in the study were added to Model I to estimate the regression model (Table 3, third column). In Model II a significant change in the coefficient of determination (R 2 change) is observed as indicated by the F-statistics (ΔR 2 = 1.9%, ΔF = 4.260, p = .002), which highlight the significant effect of the results of some of the explanatory variables considered, specifically the percentage of managers that are non-family members, the family nature of the top executive, and the presence of the founder in the management team. On the other hand, and in order to analyze if the effect of the proportion of the managers that are non-family members on the results was quadratic as opposed to linear, Model III was estimated, including the variable PNFM2. This new model did not improve in a significant way over Model II (ΔR 2 = 0.0%, ΔF = 0.092, p = .762) (Table 3, column 4). Consequently, the results of these regression analyses lead to the following conclusions of the hypotheses.
• A quadratic relationship between the proportion of non-family managers and the results of the family firms cannot be found (t = 0.304, p = .762 – Model III), although a positive linear relationship exists (t = 2.768, p = .006 – Model II). These results do not allow the acceptance of Hypothesis 1.
• A significant relationship between the homogeneity of family managers and performance does not exist (t = −0.753, p = .452 – Model II). Consequently, Hypothesis 2 cannot be accepted.
• A positive significant relationship exists between the family nature of the head of the management team and performance (t = 2.105, p = .036 – Model II), which allows to accept Hypothesis 3.
• The presence of the founder in the management team exerts a positive influence in the firm's results (t = 1.829, p = .068 – Model II), which allows to accept Hypothesis 5a but not 5b.
Finally, Model IV was estimated, which incorporates the moderating effect (Table 3). The interaction effect of a family head of the management team and the proportion of non-family managers (PNFM × HEADFAM) is negative but not significant and therefore Hypothesis 4 cannot be accepted.
Finally, the Kolmogorov–Smirnov Z-test was applied to the non-standardized residuals of Model II to test the normality of the residuals and guarantee the reliability and validity of the results. The results obtained showed the total normality of the residuals (Z of Kolmogorov–Smirnov = 1.149, p = .142).
DISCUSSION AND CONCLUSIONS
This study aims to analyze the effects that management team composition has on the results of non-listed family firms, contributing to a field of study that has been relatively neglected, as current research into TMTs rarely looks at family firms (Ling & Kellermanns, Reference Ling and Kellermanns2010; Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010). In addition, our research concentrates on non-listed family firms, a subgroup which is even less studied, amongst other reasons, because of the difficulty experienced when attempting to obtain data (Gomez-Mejia, Larraza-Kintana, & Makri, Reference Gomez-Mejia, Larraza-Kintana and Makri2003). The importance of TMTs in defining the strategic orientation of non-listed firms, and consequently in the results obtained, is very significant, more so than that of the board of directors (Dyer, Reference Dyer1986; Ward, Reference Ward1991; Neubauer & Lank, Reference Neubauer and Lank1998). On the other hand, the influence of families in the ownership and directionof the firms is more intense in those that are not listed (Cabrera-Suárez & Santana-Martin, Reference Cabrera-Suárez and Santana-Martin2004; Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010; Miller, Le Breton-Miller, & Lester, Reference Miller, Le Breton-Miller and Lester2011). Therefore, in this context, the dynamics associated with familiness and social capital are especially significant. As a consequence, it is relevant to study the influence of specific elements of the composition of the management teams such as the presence of managers that are not members of the family, and more specifically of the family nature, or not, of the top executive (Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010). Another important factor will be related with diversity within the family managers (Ling & Kellermanns Reference Ling and Kellermanns2010), and the particularly relevant role that the founding member may play (Ensley & Pearson, Reference Ensley and Pearson2005; Miller, Le Breton-Miller, & Lester, Reference Miller, Le Breton-Miller and Lester2011).
The literature suggests that the contribution of non-family managers to the management teams of family firms could have both positive and negative elements. The hypothesis presented suggests a possible inverse U relationship in the results, expecting to find evidence of the existence of a faultline in line with Minichilli, Corbetta, & MacMillan's (Reference Minichilli, Corbetta and MacMillan2010) results. However, in this study the proportion of non-family managers has a positive linear relationship with the performance. This evidence suggests that in non-listed family firms the positive contributions, in terms of cognitive diversity, professionalism and objectivity of the non-family managers overshadow the potential losses in terms of the weakening of the positive dynamics associated with familiness. Most of the firms in our study are less than 30 years old, with very few being over 40. Consequently, the conclusion is that the most substantial part of these firms is still in the first-generation stage, or in any case, that the second generation has only taken over the firm very recently. Thus, the family ties and the advantages in terms of social capital associated with this have not been eroded via a more advanced generational cousin consortium (Gersick et al., Reference Gersick, Davis, McCollom-Hampton and Lansberg1997). It is therefore possible that these firms are more capable of making use of the contributions made by non-family managers while still maintaining the united values, behaviors and orientation typically associated with family firms in their first stages of evolution (Dyer, Reference Dyer1986).
In line with what has previously been mentioned, the results obtained support the hypothesis, which suggests the positive effect of having a family member as team leader. The special characteristics of these family leaders, in terms of idiosyncratic knowledge of the firm, motivations and compromise (Cabrera-Suárez, Saá-Pérez, & García-Almeida, Reference Cabrera-Suárez, Saá-Pérez and García-Almeida2001; Miller & Le Breton Miller, Reference Miller and Le Breton-Miller2006) can translate into specific advantages in the process of decision making in management teams and consequently on decision quality. In fact, it was not possible to find significant results in relation to the hypothesis regarding the negative effect that the family nature of the leadership of the firm may have on the motivation and involvement of managers who are non-family members.
In cases where the founding members were the team leaders, a significant and positive influence on the performance of the family firm was observed. Consequently, in the non-listed family firms, the presence of the founder has positive effects on the dynamics within the management teams, thanks to the psychological feeling of belonging, stewardship motivation and reciprocal altruism that the founder is capable of generating (Dyer, Reference Dyer1986; Karra, Tracey, & Phillips, Reference Karra, Tracey and Phillips2006). Similarly, in these firms, the possible inconveniences derived from a high concentration of the power in the hands of the founder and the problems of self-control highlighted, amongst others, by Gedajlovic, Lubatkin, and Schulze (Reference Gedajlovic, Lubatkin and Schulze2004) and Buyl et al. (Reference Buyl, Boone, Hendriks and Matthyssens2011) do not succeed in cancelling out the positive effects previously mentioned.
The existing diversity within the family managers was also considered. More specifically, the assumption was that the higher the proportion of siblings compared to the total number of family managers, the more homogeneous the family managers would be. In turn, the expectation was that a greater homogeneity would have positive effects on the performance. However, the results obtained point toward a negative influence, though not significant. The explanation for this unexpected result can be that a high homogeneity between family members would lead to phenomena associated with group thought, which would derive in a vision that was excessively uniform and limited regarding the needs and strategic options of the firm, worsening the firm's performance. Another alternative explanation is that the variable used does not represent the variability within the family executives adequately, as differences in approaches, values and expectations may not be determined by sibling status, but other variables such as age, birth order or even sex. The literature states that these variables influence the processes of socialization and development of members of family firms, and consequently in their differing attitudes and visions with respect to the family firm (e.g., Kets de Vries, Reference Kets de Vries1985; Handler, Reference Handler1989; Friedman, Reference Friedman1991; Dumas, Reference Dumas1992; Lansberg, Reference Lansberg1999).
Therefore, any future investigation in this line of study should include, as demographic variables specifically in the study of the TMTs of family firms, variables such as degree of relatedness, age differences and birth order of the family managers. The analysis of these variables would contribute to the body of knowledge that has accumulated in the general literature about TMTs focused on the study of the influence of demographic variables on the functioning of management teams. Increasing the number of databases in order to increase the number and variability of the firms included in terms of generational evolution, size and geographic location, amongst others, would also be worthwhile.
A deeper analysis of the specificities in the TMTs that affect their capacity to behave in an integrated manner and improve their efficiency is another future line of investigation. The results highlight the importance in differentiating between family and non-family members in management teams, as these can have differences in their motivations, attitudes and intentions. Consequently, the expectation is that the adoption of a perspective of social capital in relation to the question of familiness will enable a better comprehension of the dynamics between the members of top managers in family firms as other authors have mentioned (e.g., Ensley & Pearson, Reference Ensley and Pearson2005; Minichilli, Corbetta, & MacMillan, Reference Minichilli, Corbetta and MacMillan2010).
Thus, future research should adopt methodological strategies that allow the researchers to obtain primary data, and of a longitudinal character, allowing the questions to be adequately assessed. The design of questionnaires and in-depth interviews directed to members of the TMTs could ease the collection of data of variables that are not readily accessible via secondary sources.
Finally, regarding the practical implications of this study for non-listed family firms, the conclusion is that though family leadership increases value to the TMT of these firms, the introduction of managers that are not members of the family creates a more relevant positive influence on performance. Consequently, these firms should develop human resource policies in a way that is capable of attracting, retaining and integrating this human capital in a way that maximizes the synergy with the elements contributed by the family managers – compromise, values and idiosyncratic knowledge – and those contributed by non-family members – new ideas, objectivity and a wider range of experience. Particularly relevant is the establishment of policies that prevent non-family members of the management team from feeling that family leadership is an obstacle for their professional development and that their capabilities and contributions will receive fair acknowledgement.
Acknowledgements
This investigation has received financial support from the Spanish Ministry of Science and Innovation (Project ECO2008-00265/ECON).
